Last night, I was on the phone with my mother trying to explain the problems I was having with my research. Right now, I am doing uncertainty analysis and I was getting results whose error was seven orders of magnitude greater than the actual data value. Clearly this is wrong.
So I used the following analogy to get across the problem to my mom: “It is like thinking you have one M&M, you know, plus or minus 10 million M&Ms.”
Mom: “Yikes! That is a lot of variance!”
Me: “Yup…And what I’d really like is one M&M plus or minus half an M&M.”
Mom: “So, give or take an ‘M’.”
I thought it was cute. I wonder if this is a good way to teach error analysis. It worked seamlessly here…